/
compute_and_store_expectation_value.py
executable file
·176 lines (148 loc) · 6.67 KB
/
compute_and_store_expectation_value.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
#!/usr/bin/env python
import sys
import glob
import pickle
import argparse
from pathlib import Path
import numpy as np
import pandas as pd
import qiskit
import qcopt
def parse_args():
parser = argparse.ArgumentParser()
parser.add_argument("-p", "--path", type=str, default=None, help="path to project")
parser.add_argument("-P", type=int, default=1, help="P-value for QAOA")
parser.add_argument("-N", type=int, default=20, help="Number of nodes")
parser.add_argument(
"--graphtype", type=str, default=None, help="Graph type to load (d3, p20, p50, p80)"
)
parser.add_argument(
"--graphname", type=str, default=None, help="Graph name (1, 2, 1 to 10, 11 to 30, ....)"
)
parser.add_argument(
"--repname", type=str, default=None, help="Rep name (1, 1 to 5, extra 1, extra 1 to 5, ...)"
)
parser.add_argument(
"--threads", type=int, default=2, help="Number of threads in circuit simulation"
)
args = parser.parse_args()
return args
def get_pickle(ROOT, args, N, graph_name, rep_glob, verbose=0):
retval = []
base_path = f"{ROOT}/benchmark_results/QAOA+_P{args.P}_qasm/N{N}_{args.graphtype}_graphs/{graph_name}/{rep_glob}"
pickle_files = glob.glob(base_path)
for pklfile in pickle_files:
if verbose:
print("Loading pickle file:", pklfile)
with open(pklfile, "rb") as pf:
res = pickle.load(pf)
return res
def get_output_dist_dict(probs, G):
output_dict = {
hammingweight: {"valid": 0, "invalid": 0} for hammingweight in range(len(G.nodes) + 1)
}
for bitstring, p in probs.items():
if qcopt.graph_funcs.is_indset(bitstring, G):
output_dict[int(qcopt.helper_funcs.hamming_weight(bitstring))]["valid"] += p
else:
output_dict[int(qcopt.helper_funcs.hamming_weight(bitstring))]["invalid"] += p
return output_dict
def main():
args = parse_args()
ROOT = args.path
if ROOT[-1] != "/":
ROOT += "/"
sys.path.append(ROOT)
# Parse input params
csv_savepath = (
ROOT + f"benchmark_results/QAOA+_expectation_values/N{args.N}_{args.graphtype}_graphs/"
)
Path(csv_savepath).mkdir(parents=True, exist_ok=True)
csv_savename = f"qaoa+_P{args.P}_{args.graphtype}_graphs_{'_'.join(args.graphname.split())}_rep{'_'.join(args.repname.split())}.csv"
# Only allow a single p and graph_type, but multiple graph_names and reps
if "to" in args.graphname:
lowerlim = int(args.graphname.split()[0])
upperlim = int(args.graphname.split()[-1])
graph_names = [f"G{i}" for i in np.arange(lowerlim, upperlim + 1)]
else:
graph_names = [f"G{args.graphname}"]
if "extra" in args.repname:
if "to" in args.repname:
lowerlim = int(args.repname.split()[1])
upperlim = int(args.repname.split()[-1])
rep_globs = [f"extra_lambda/*rep{i}.pickle" for i in np.arange(lowerlim, upperlim + 1)]
else:
rep_globs = [f"extra_lambda/*rep{args.repname.split()[-1]}.pickle"]
else:
if "to" in args.repname:
lowerlim = int(args.repname.split()[0])
upperlim = int(args.repname.split()[-1])
rep_globs = [f"*rep{i}.pickle" for i in np.arange(lowerlim, upperlim + 1)]
else:
rep_globs = [f"*rep{args.repname}.pickle"]
backend = qiskit.Aer.get_backend("aer_simulator_statevector", max_parallel_threads=args.threads)
columns = ["lambda", "expectation_value", "p", "N", "graph_type", "graph_name", "rep_name"]
df = pd.DataFrame(columns=columns)
for graph_name in graph_names:
for rep_glob in rep_globs:
qaoa_plus_data = get_pickle(ROOT, args, args.N, graph_name, rep_glob, verbose=0)
if "extra" in rep_glob:
cur_rep_name = f"extra_{rep_glob.split('*')[-1].strip('.pickle')}"
else:
cur_rep_name = rep_glob.split("*")[-1].strip(".pickle")
print(
f"Processing QAOA+ p = {args.P}, {args.graphtype} graph {graph_name} {cur_rep_name}"
)
dist_savepath = (
ROOT
+ f"benchmark_results/QAOA+_output_distributions/P{args.P}_{args.graphtype}_N{args.N}/{graph_name}_{cur_rep_name}/"
)
Path(dist_savepath).mkdir(parents=True, exist_ok=True)
for data_dict in qaoa_plus_data:
rounded_lambda = round(data_dict["lambda"], 3)
graph_fn = "/".join(data_dict["graph"].split("/")[-3:])
G = qcopt.graph_funcs.graph_from_file(ROOT + graph_fn)
# Evaluate the optimized circuit
print(f"\tSimulating circuit, lambda = {rounded_lambda}")
circ = qcopt.ansatz.qaoa_plus.construct_qaoa_plus(
args.P, G, data_dict["opt_params"], measure=False
)
circ.save_statevector()
result = qiskit.execute(circ, backend=backend).result()
probs = qiskit.quantum_info.Statevector(
result.get_statevector(circ)
).probabilities_dict(decimals=7)
# Compute the expectation of the objective function
print(f"\t\tComputing expectation value...")
expected_energy = qcopt.qaoa_plus_mis.expectation_value(probs, G, rounded_lambda)
# Save to the DataFrame
save_values = [
rounded_lambda,
expected_energy,
args.P,
args.N,
args.graphtype,
graph_name,
cur_rep_name,
]
df = pd.concat([df, pd.DataFrame([save_values], columns=columns)])
# Analyze the valid/invalid output states in the final distribution
print(f"\t\tAnalyzing output distribution...")
output_dist_dict = get_output_dist_dict(probs, G)
# Save to pickle
with open(
dist_savepath + f"hamming_histogram_lambda_{rounded_lambda}.pickle", "wb"
) as pf:
pickle.dump(output_dist_dict, pf)
# Save the top 100 most probable bitstrings to pickle
top_strings = sorted(
[(key, val) for key, val in probs.items()], key=lambda p: p[1], reverse=True
)[:100]
with open(
dist_savepath + f"top_100strings_lambda_{rounded_lambda}.pickle", "wb"
) as pf:
pickle.dump(top_strings, pf)
# Save the DataFrame as csv
df.to_csv(csv_savepath + csv_savename)
if __name__ == "__main__":
main()